# Data science 5 SBU — Data

来源：https://www.kaggle.com/competitions/data-science-5-sbu/data

采集时间：2026-10-19T10:04:25Z


### Files

* **`train.csv`** - The primary dataset for training your machine learning models. It contains the features and the target variable (`Policy Cost`) for model learning.
* **`x_test.csv`** - This is the test set on which you will make predictions. It contains the same features as `train.csv` but *without* the `Policy Cost` column. Your final submission file should contain predictions for the `Policy Cost` for each `ID` in this file.

* **`sample_submission.csv`** - A sample submission file provided in the correct format. This file shows you exactly how your submission should be structured.

### Columns

Here's a detailed description of the columns you will encounter across the datasets:

#### `train.csv` and `x_test.csv` (Feature Columns)

These files share the same feature columns. Note that the original column names have been anonymized for the competition.

* `ID` - A unique numerical identifier for each entry/row in the dataset.
* `Years Lived` - Anonymized representation of the customer's age.
* `Sex` - Anonymized representation of the customer's gender (e.g., 'Man', 'Woman').
* `Yearly Earnings` - Anonymized representation of the customer's annual income.
* `Relationship Status` - Anonymized representation of the customer's marital status (e.g., 'Unmarried', 'Spouse', 'Separated').
* `Dependent Count` - Anonymized representation of the number of dependents the customer has.
* `Academic Standing` - Anonymized representation of the customer's education level (e.g., 'Undergraduate', 'Graduate', 'Doctorate', 'Secondary').
* `Job Title` - Anonymized representation of the customer's occupation (e.g., 'Working', 'Freelancer', 'Jobless').
* `Wellness Index` - Anonymized representation of the customer's health score.
* `Region` - Anonymized representation of the customer's geographical location (e.g., 'City', 'Exurban', 'Countryside').
* `Coverage Class` - Anonymized representation of the policy type (e.g., 'Superior', 'Full', 'Standard').
* `Prior Claims` - Anonymized representation of the number of previous claims made by the customer.
* `Automobile Age` - Anonymized representation of the age of the customer's vehicle.
* `Financial Rating` - Anonymized representation of the customer's credit score.
* `Coverage Period` - Anonymized representation of the duration of the insurance policy.
* `Coverage Commencement` - Anonymized representation of the policy start date.
* `Client Review` - Anonymized representation of customer feedback (e.g., 'Satisfactory', 'Moderate', 'Unsatisfactory').
* `Tobacco Use` - Anonymized representation of the customer's smoking status (e.g., 'Smoker', 'Non-Smoker').
* `Physical Activity` - Anonymized representation of the customer's exercise frequency (e.g., 'Everyday', 'Once a Week', 'Once a Month', 'Seldom').
* `Asset Category` - Anonymized representation of the property type (e.g., 'Residence', 'Flat', 'Unit').
* `Policy Cost` - **(ONLY in `train.csv`)** The target variable you need to predict. This is the premium amount for the insurance policy.

* `ID` - A unique numerical identifier for each entry, matching the `ID` in `x_test.csv`.
* `Policy Cost` - The actual premium amount for the corresponding `ID` in `x_test.csv`.

#### `sample_submission.csv` (Submission Format)

Your submission file should contain two columns:

* `ID` - The unique identifier from `x_test.csv`.
* `Policy Cost` - Your predicted `Policy Cost` for that `ID`.